Embodied AI for Autonomous Medical Robotics

Institut

Chair for Computer Aided Medical Procedures & Augmented Reality (TUM-CIT)

Typ

Semesterarbeit / Masterarbeit /

Inhalt

experimentell / theoretisch / konstruktiv /  

Beschreibung

Background and Motivation:

Embodied AI and robot learning are rapidly transforming how robots perceive, reason, and act in complex real-world environments. In medical settings—such as ultrasound scanning and microsurgical manipulation—robots need go beyond image-level understanding and achieve actionable spatial reasoning and closed-loop control: interpreting multimodal context, planning safe and effective motions, and adapting actions in real time under strict precision constraints. To push medical robotics toward greater autonomy and stronger clinical relevance, we are seeking highly motivated students to work on real robotic platforms in Munich, aiming to publish solid, publishable contributions in top-tier robotics journals (e.g., IEEE T-RO, IEEE TASE, T-Mech, RA-L).

 

Research Objectives:

Applicants may choose one of the following directions:

Track A: VLA / Robot Learning for Autonomous Ultrasound Robotics

• Develop VLA or state-of-the-art robot learning methods for autonomous ultrasound scanning

• Enable multimodal understanding, probe motion planning, interactive scanning, and robust control

• Validate the system through experiments on real robotic ultrasound platforms

 

Track B: LLM-Agent for Microsurgical Robot Control & Real-Time Image Processing

• Build LLM-based agents integrated with real-time microscopic medical image analysis

• Study multimodal reasoning, decision-making, and closed-loop control in high-precision microsurgical tasks

• Evaluate the system on real robotic platforms for fine-grained manipulation

 

Voraussetzungen

 

Prerequisites:

• Strong background in robotics (e.g., kinematics, dynamics, ROS/ROS2), ideally with hands-on hardware experience(KUKA or Franka)

• Excellent programming skills in Python/C++

• Strong interest in robot learning, multimodal AI, or medical robotics

• Availability for on-site work in Munich

 

What We Offer:

• Hands-on work with real robots

• A rigorous and academically oriented research environment

• Excellent preparation for a PhD in medical robotics or embodied AI

 

Application:

Interested applicants, please email: a). Brief CV b) Transcripts of TUM to: 

ning.wang@tum.de

Möglicher Beginn

sofort

Kontakt

Xuesong Li
Xuesong.Litum.de